Re: Does entity integrity imply entity identity?

From: Keith H Duggar <duggar_at_alum.mit.edu>
Date: Thu, 6 Aug 2009 07:44:29 -0700 (PDT)
Message-ID: <bdf0f71a-6a5f-4644-a1cd-1321c85178aa_at_k30g2000yqf.googlegroups.com>


On Aug 6, 3:57 am, "Mr. Scott" <do_not_re..._at_noone.com> wrote:
> "Bob Badour" <bbad..._at_pei.sympatico.ca> wrote in message
> > Mr. Scott wrote:
> >> "Bob Badour" <bbad..._at_pei.sympatico.ca> wrote in message
> >> If the requirement is to record in the database as much information as is
> >> available, then how can the fact that a value is at present unknown be
> >> recorded?
>
> > By stating so in the database. That's what relations do: make statements.
>
> How, exactly? If you have columns K, A, B, and C such that for a given K it
> is permissible for the values for A, B or C to be at present unknown, but
> not both A and B, how would you lay out the tables and referential
> constraints without using nulls?

I'm curious:

  1. Do you think this question of yours is *new*? How many times would you guess this same question has been asked in this forum?
  2. Did you read any of papers Bob alluded to? What about books on basic theory that cover the subject of nulls and alternative solutions?
  3. Did you search the cdt group archives for related posts? What did you find so far?
  4. Is your goal to thoroughly learn? Are you willing to put work into that goal?
  5. Are you too lazy to research the subject? Are you looking to be spoon-fed?
  6. Do you have another goal such as to troll or to entertain yourself?

Answering the above will help the cdt folks tailor their answers appropriately.

KHD Received on Thu Aug 06 2009 - 16:44:29 CEST

Original text of this message